2.1 Changes in phytoplankton density and water quality following water diversion
Water diversion has the potential to change the phytoplankton community composition12 and improve the water quality of eutrophic lakes10. Because extraordinary disturbances occurred with the operation of the South-to-North Water Diversion Project in Luoma, Nansi and Dongping lakes since October 2013, phytoplankton communities in the three lakes were affected in various ways. Sharp decreases in algal density, particularly for cyanobacteria, have been shown in 2014 in the three lakes. Thereafter algal density remained almost unchanged during the many years of water diversion in the studied lakes (Fig. 1 and Fig. S1). Comparing these lakes, after water diversion the algal population decreased more in Nansi and Dongping lakes than in Luoma Lake. The annually averaged total algal density changed from 10.9 and 8.7 million cells/L in Nansi and Dongping lakes before water diversion to respectively 4.5 and 4.1 million cells/L after water diversion, while the annually averaged density of total algae only decreased in 2014, and then slightly increased in Luoma Lake. A possible reason lies in the different concentrations of nutrients in the various lakes: the average concentration of TN after water diversion was greater in Luoma Lake (1.58 mg/L) than in the other lakes (0.84 mg/L in Dongping Lake and 0.91 mg/L in Nansi Lake) (Fig. 2 and Fig. S2). The other reason could be that Luoma Lake (located in Jiangsu Province) had a more suitable temperature and climatic conditions than the other two lakes, causing the cyanobacteria to have a stronger resistance to interference from water diversion.
The concentration of nutrients decreased after water diversion in the three lakes. The average concentration of TN in Luoma Lake was 2.58 mg/L before the water transfer, while the concentrations were 1.28 mg/L in Nansi Lake and 1.92 mg/L in Dongping Lake. A tremendous decline in the concentrations of TN was observed after the disturbances of South-to-North Water Diversion Project in Luoma Lake (1.59 mg/L), Nansi Lake (0.92 mg/L) and Dongping Lake (0.84 mg/L) (Fig. 2 and Fig. S2). Before the water diversion, a higher average concentration of TP was detected in Nansi Lake (0.08mg/L) compared with Luoma Lake (0.03 mg/L) and Dongping Lake (0.04 mg/L). However, the average concentration of TP tended to be consistent (0.04 mg/L) across the three lakes for the 6-year period of water importation (Fig. 2 and Fig. S2). The main reasons might be that pollution-control projects and dozens of associated measures were implemented, including sewage diversion projects, sewage treatment plant construction, wetland ecological restoration, and closing down of factories with substandard emissions around the lake30.
Phosphorus is needed to maintain cell and membrane structure, energy metabolism, the synthesis and expression of genetic material, and a variety of other metabolic and regulatory processes. A low level of accessible P in the environment can compromise algae, causing decreased growth31. Currently, lakes are often said to be primarily phosphorus-limited31; however, counter-examples cast doubt in some researchers’ minds32. Herein we compared TP in the three lakes, the concentrations of TP manifested minor changes, with the average concentration (0.04 mg/L) always being low in Dongping Lake, while more variation occurred in Nansi Lake before and after water transfer. The phytoplankton density in Dongping Lake was consistent with that in Nansi Lake during the study period, but, on the other hand, the algal density did not change in Luoma Lake, where phosphorus concentration clearly changed in stage two. The relation between TP and algal density in different lakes indicates that phosphorus was not the limiting factor in Luoma and Dongping lakes, which are mesotrophic–eutrophic lakes33, which supports the concept that the “phosphorus-limitation paradigm” applies only to oligotrophic lakes.
Nutrient management focused on a single nutrient is likely to result in reductions in multiple nutrients. Thus it is often not possible to credit phytoplankton changes only to control of a single nutrient34. Nitrogen (N) and phosphorus (P) are the key nutrients for phytoplankton: the combined effects of N and P enrichment have accelerated eutrophication and proliferation of phytoplankton on a global scale, particularly for harmful cyanobacteria35. Herein we found that TN and TP quickly decrease in stage one of water diversion, and the algal density also simultaneously reduced in Nansi and Dongping lakes. However, in Luoma Lake a decrease of TN and no great variation of TP occurred, and the phytoplankton exhibited no obvious decrease (Fig. 1-2 and Fig. S1-S2). Traditional approaches for cyanobacterial control have focused on reducing phosphorus (P) inputs to impaired freshwater systems, based on the fact that excessive P relative to N inputs (or low N-to-P ratios) were linked to a tendency for systems to be dominated by cyanobacteria36. But our results indicate that reducing both N and P in a water body provides the best opportunity to reduce phytoplankton density.
Excessive growth of phytoplankton, particularly in the form of cyanobacterial blooms, has become a growing global problem37–38. Many species of cyanobacteria are capable of producing toxic and noxious compounds that degrade water quality and pose a risk to human and aquatic ecosystem health39. Previous researchers had found warmer waters to favour the incidence of cyanobacterial blooms in many lakes40. However, the average annual water temperature had no practically significant changes in the three lakes during the nine years’ monitoring (Fig. 2 and Fig. S2). The cyanobacterial density clearly decreased in Nansi and Dongping lakes, while only weak variation occurred in Luoma Lake. Meanwhile the frequency of cyanobacterial blooms did not increase in the three lakes. Overall the intensity of cyanobacterial blooms in these water-diversion lakes did not increase under the influence of global warming.
2.2 Influence of environmental selection on abundant and rare phytoplankton sub-communities
It has been suggested that rare and abundant taxa of microorganisms may have different roles in natural habitats41. Abundant species usually contribute a significant role in the ecosystem and participate in major biogeochemical processes such as carbon ﬂow and processing of nutrients, whereas rare taxa can be regarded as propagule banks and play minor but non-negligible roles41,42. A ubiquitous pattern has been observed across ecosystems whereby many species are rare, and a few species are abundant20-21, 23-24. This pattern is also found in our research: there was no species classified as always abundant in 2015 in Luoma Lake, and the number of conditionally abundant species varied from 1 (0.4%) in 2018 to 8 (4.0%) in 2015. Conditionally rare species made up a greater proportion in Luoma Lake, reaching 77.4% in 2016, and we found that the fewest conditionally abundant and rare species were identified in 2016, with 47 species (20.1%), whereas the number of such species peaked at 67 (30%) in 2019. However, there were no always rare species nor moderately abundant taxa classified in Luoma Lake (Table S2). As far as Nansi Lake is concerned, in 2015 there were 6 (2.8%) algal species that were considered to be abundant, which was the minimum number, while the maximum number of abundant algal species was 22 (12.0%), which occurred in 2014. The smallest (116 (68.6%)) and largest (168 (72.4%)) numbers of rare species were classified in 2010 and 2017, respectively. The identified numbers and proportions of conditionally abundant and rare taxa were distinct in different years, and no moderately abundant taxa nor always rare taxa were found during the study period in Nansi Lake (Table S3). The number and proportion of abundant species varied from 1 (0.7%) in 2011 to 10 (6.6%) in 2019 in Dongping Lake, and no always abundant algae were present in 2010 or 2011 (Table S4). In addition, the number and proportion of rare species varied from 85 (56.3%) in 2019 to 128 (67.0%) in 2017. There were fewer conditionally rare taxa in Dongping Lake compared with Luoma and Nansi lakes. Consistent with the results in Luoma and Nansi lakes, the number and proportion of conditionally rare and abundant taxa varied dramatically in different sampling years, and no always rare species nor moderately abundant taxa were identified in Dongping Lake (Table S2- S4).
The rare taxa were observed to be dissimilar and to encompass many more unique species in the three lakes. The rare sub-communities could be inert in one specific environment20-21, but may become abundant taxa upon environmental changes22–23. The rare species underwent dramatic changes in Luoma, Nansi and Dongping lakes following from the water diversion project. Scenedesmus obtusus var. apiculatus was converted from a rare species in Luoma Lake into abundant taxa following water diversion. Other rare species (Cyclotella meneghiniana, Pedinomonas minor, and Achnanthes exigua) were transformed into abundant algae after water diversion in Nansi Lake. Meanwhile, Chlamydomonas microsphaera and Chlorella ellipsoidea were rare taxa before water diversion in Dongping Lake, but they became abundant species after water transfer.
The effects of water diversion on abundant species variation in the three lakes were similar. The number of abundant algae in the three lakes increased in 2013 and 2014 (Table S2- S4), the reasons for this phenomenon might be the dramatic changes of water quality in the three lakes due to water importation, causing rare species and conditionally abundant and rare taxa to be transformed into abundant algae.
Resistance occurs when some members in a microbial community show high tolerance to environmental disturbances43. Chlorella vulgaris, Chroomonas acuta and Synedra acus remained classified as always abundant algae under the pressure of the water diversion project and had stronger environmental tolerance than other abundant species (Table S5- S7). Distinct responses of rare and abundant taxa to disturbances might arise from their different life-history strategies. The number of abundant algae remained subsequently unchanged during stage two, which suggests that the phytoplankton communities — especially for abundant species — had a degree of resilience. After a disturbance, the altered microbial community could rapidly recover to its original state or establish an alternative stable state44–45.
Phytoplankton are considered to be an indicator of environmental change and ecosystem state owing to their quick and strong responses to environmental disturbances46. At the same time, environmental factors such as temperature, light, or nutrient concentrations are known to affect abundance and activity of microbial taxa47–48. We adopted Mantel tests to elucidate the relationship between all, abundant, or rare phytoplankton and environmental factors in the three lakes. Studies have demonstrated that different controlling factors constrain the structures of the rare and abundant sub-communities23. Herein we observe statistically significant (p < 0.05) positive responses between all or abundant taxa and WT; however, there was no statistically significant relationship detected between rare species and WT (Table 1), which could be ascribed to the fact that these species may have different growth and activity and thus respond differently to the environmental variable. The result differed from a previous report that temperature showed the strongest correlation with rare species, which is mainly because diatom blooms occurred in the previously studied water body49. In addition, phytoplankton had different responses to pH, transparency, DO, CODMn and TP in the three lakes (Table 1), which are related to the environment of the respective lake itself.
2.3 Reestablishment and stability of the phytoplankton community structure
Phytoplankton enable fast responses to abrupt water fluctuation in lakes, and phytoplankton community composition stability and diversity could rapidly change14. Naselli-Flores and Barone (2000)50 reported that the unique hydraulic regimes, rather than nutrient availability, were the main factors affecting the algal community structure in lakes characterized by apparent water mixing. Additionally, some long-term ﬁeld data for a shallow ﬂoodplain lake indicated the relevance of water ﬂuctuations in driving the shift from a state dominated by free-ﬂoating plants to a state dominated by phytoplankton51.
When the stability of the networks was assessed by calculating the ratio between density (D) and clustering coefﬁcient (transitivity, T) of the network53, the values of D/T were smaller during non-water-diversion months than water-diversion months, which meant the stability of the algal community structure composition was greater (Fig. 3 and Table S8). Herein it is apparent that there was more stability in non-water-diversion months than in water-diversion months, and the water diversion project acted as an external disturbance to the algal communities of the receiving lakes and, as such, the project was able to directly inﬂuence the stability of the phytoplankton community structure.
In addition, co-occurrence networks were constructed in every year depending on whether or not water was transferred. The value of D/T was lower in non-water-diversion months than in water-diversion months during the period of 2013–2016 in Luoma Lake, while the opposite result occurred in 2017–2019 (Fig. S3, Table S9). Moreover, in Dongping Lake D/T was lower in non-water-diversion months than during water-diversion months from 2014 to 2016, while similar D/T values were calculated for non-water-diversion months and water-diversion months during 2017–2019 (Fig. S5, Table S11). However, in Nansi Lake the value of D/T was always lower in non-water-diversion months than during water-diversion months (Fig. S4, Table S10). Hence we classified different water diversion phases: before water diversion (2011−2013), stage one of water transfer (2014−2016), and stage two of water importation (2017–2019).
The stability of the algal communities in Luoma and Dongping lakes weakened during phase one of water transfer, while there were no stability changes in Nansi Lake. As the long-term planned water diversion continued the stability in Luoma and Dongping lakes recovered, whilst it decreased in Nansi Lake (Fig. 4 and Fig. S6). The difference in stability changes was mainly caused by different lakes’ morphologies: Luoma and Dongping lakes are nearly circular lakes, but Nansi Lake is shaped like a ribbon and behaves more like a river (Fig. S7); another reason could be that the number of rivers flowing into these lakes are different —only one main river (Dawenhe River) flows into Dongping Lake and two main rivers (Yi River and Zhong Canal) flow into Luoma Lake, while 53 rivers flow into Nansi Lake, so that more disturbed locations existed in Nansi Lake than the other studied lakes.
Some systems may respond smoothly to environmental changes, whereas others may show threshold responses, responding abruptly only beyond a critical value of disturbance52. Surprisingly, the altered microbial communities re-established an alternative stable state in Luoma and Dongping lakes, but the stability reduced in Nansi Lake in stage two of water diversion. The results indicate that the phytoplankton in circular lakes had quicker responses and recoveries to the water transfer than phytoplankton in strip-shaped lakes. We guess that the stability of the phytoplankton community in Nansi Lake will also eventually recover in the future with continuation of the water transfer process, and we will conduct persistent monitoring and observation on the three lakes to check this.
Phytoplankton diversity in the lakes includes a range of functional groups, which play critical roles in transferring energy across the food web14–15. Therefore, tracking changes in the structure and diversity of algal groups is the key to characterising ecosystem function and measuring the impacts of disturbances. Despite the diversity of algal communities in the three lakes, there were certain consistent changes following water importation: surprisingly, the diversity and evenness of all, abundant and rare algae exhibited no dramatic changes in the three lakes after the disturbance of water diversion (Fig. 5, Fig. S8- S9). This result was not consistent with the studies of Yang et al.54 and Dai et al.55, where phytoplankton community diversity and evenness index were increased in Gonghu Bay when it was affected by water diversion. Importantly, their finding related to the effect of water diversion on community diversity mainly focused on short-term changes (i.e. over several months). Moreover, we can guess that the change in disturbed community diversity was affected by the sensitivity of the respective community to environmental change.
Whether the community structure would change irreversibly in response to disturbance is a current concern that needs to be solved. Tromas et al. viewed cyanobacterial blooms as a biological disturbance, quantified by their impact on the surrounding microbial community, and found that the community changes cyclically over the course of a year, with a repeatable pattern from year to year56. In our study, the algal composition became more and more similar to that in 2014 (the state directly after water diversion) during the long-term period of observation in the three lakes (Fig. 6, Fig. S10- S11). This could reflect resilience that allows the altered microbial community to recover to its original state or establish an alternative state after a disturbance44–45. With different consequences caused by disturbance, community structures affected by algal blooms are restored to their original status after a short period, while the communities affected by water diversion only stabilize to a new state; the differences in the recovery time and state could depend upon differences in the disturbance intensity and type.
The community compositions, particularly for abundant algae, were dissimilar in the three lakes, the main contributions to these distinct compositions in 2011 were the high abundances of Pseudanabaena limnetica (42.5%), Chroomonas acuta (30.7%) and Chlorella vulgaris (24.0%) in respectively Luoma, Nansi and Dongping lakes; in 2012 Pseudanabaena limnetica (54.0% in Luoma lake and 19.3% in Nansi lake) and Chroomonas acuta (27.9% in Dongping lake) were dominant algal species in different lakes; and in 2013 Pseudanabaena limnetica (34.8%), Chlorella vulgaris (24.0%) and Chroomonas acuta (30.7%) had the highest abundances in Luoma, Nansi and Dongping lakes, respectively. Meanwhile, no species homogenization across the lakes occurred after water diversion. The community compositions were distinct and there were different proportions of dominant algal species after water importation in the three lakes (Tables S2–S7). Unsurprisingly, this phenomenon was also found between two different lakes that were connected by water transfer (Fig. S10- S11). Contradictory conclusions have been reported that water diversion projects would lead to a massive introduction of distinct biodiversity in lakes, bio-invasion of lakes, and biotic homogenization across lakes12. Guo et al. investigated the influence of a water diversion project on fish communities and identified that increased hydrological connectivity had potentially created an “invasion highway” and promoted a potential biotic homogenization among the impounded lakes29. In a more comprehensive investigation, a wide variety of aquatic organisms (molluscs, zooplankton, crustaceans, insects, ﬁsh, amphibians, reptiles, mammals, and plants) was studied and the fact that a water diversion project caused homogenization between lakes was similarly proved57. Previous research has mainly focused on the effect of water transfer on aquatic plants and animals. Compared with plants and animals, diverted phytoplankton had more sensitive responses to environmental change (e.g. the change in nutrient levels within the three lakes), and our finding eliminated the concern that water diversion can cause phytoplankton homogeneity across different lakes.